agents-libagents-lib
AgentsMCP ServersWorkflowsTutorialsComparisonsTemplatesNewsAbout
MU

Muse

⭐ 0•Multi-language•Consumer Agent

Meta's personal AI agent — persistent assistant across apps, calendar, and conversations with on-device memory.

Website →
ConsumerPersonal AssistantMetaPersistentOn-DeviceCross-Platform

Overview

Muse is Meta's consumer personal AI agent, launched September 8, 2026 (659 HN points). Positioned alongside Google Gemini's personal assistant, Microsoft Copilot 365, and Apple's iOS 27 Siri. Muse acts as a persistent assistant across the user's apps, calendar, and conversations, with on-device memory and a provider-agnostic backend. Available on iOS, Android, Web, and desktop, with a free tier and paid Muse+ subscriptions.

Features

  • ✓Persistent assistant across apps, calendar, conversations, and email
  • ✓On-device memory with user-controlled privacy boundaries
  • ✓Provider-agnostic backend: Meta models plus third-party providers
  • ✓Proactive suggestions and follow-up synthesis
  • ✓Cross-platform: iOS, Android, Web, macOS, Windows
  • ✓Voice + text interfaces
  • ✓Free tier with usage limits and Muse+ subscriptions

Installation

Download from the App Store, Google Play, or visit muse.meta.ai

Pros

  • +Polished consumer UX built for non-technical users
  • +Meta ecosystem integrations (WhatsApp, Messenger, Facebook, Instagram)
  • +On-device memory with privacy controls
  • +Free tier lowers the barrier to entry
  • +Provider-agnostic; not locked to a single model

Cons

  • −Requires Meta account and associated data
  • −Less powerful than dedicated agent frameworks for complex workflows
  • −Limited third-party tool access compared to MCP-first agents
  • −Privacy concerns around Meta's data practices
  • −No code/API surface for developer extensibility

Alternatives

Google Gemini →Microsoft Copilot 365 →Apple Siri (iOS 27) →Claude Cowork →

Documentation

Muse

Overview

Muse is Meta's personal AI agent, launched on September 8, 2026 (659 HN points — one of the highest-scoring consumer agent launches of the period). Muse is Meta's entry into the consumer personal-agent market, positioned alongside Google Gemini's personal assistant ambitions, Microsoft Copilot 365, and Apple's iOS 27 Siri.

Where Claude Cowork targets knowledge workers and coding agents target developers, Muse is aimed at the general consumer. It acts as a persistent assistant across a user's apps, calendar, and conversations, with on-device memory and a provider-agnostic backend.

The launch itself became a cultural moment when a secondary story — Muse, the band, lost its social media handles to Muse, Meta's new AI agent (184 points) — made the round, highlighting the consumer-agent naming collision problem.

Features

  • Persistent personal assistant: Operates across apps, calendar, conversations, and emails
  • On-device memory: User context stored locally with optional cloud sync
  • Provider-agnostic backend: Runs on Meta's in-house models plus third-party providers
  • Proactive suggestions: Surfaces reminders, follow-ups, and synthesized updates
  • Cross-platform: iOS, Android, Web, and desktop clients
  • Privacy posture: User-controlled data boundaries with explicit permission model
  • Voice + text interfaces: Conversational and quick-action modes

Availability

  • Mobile: iOS and Android (App Store / Google Play)
  • Web: muse.meta.ai
  • Desktop: macOS and Windows apps
  • Free tier: Available at no cost with usage limits
  • Muse+: Subscriptions with higher limits and priority features

Use Cases

  • Morning briefings: Daily summary of calendar, messages, weather, and tasks
  • Conversation recall: Search across prior chat threads and emails for facts
  • Action delegation: Schedule meetings, send messages, file travel plans
  • Content synthesis: Summarize long threads, emails, and documents
  • Cross-app orchestration: Chain actions across calendar, email, contacts, and notes

Pros

  • ✅ Polished consumer UX — built for non-technical users
  • ✅ Meta ecosystem integrations (WhatsApp, Messenger, Facebook, Instagram)
  • ✅ On-device memory with privacy controls
  • ✅ Free tier lowers the barrier to entry
  • ✅ Provider-agnostic; not locked to a single model

Cons

  • ❌ Requires Meta account (and associated data)
  • ❌ Less powerful than dedicated agent frameworks for complex workflows
  • ❌ Limited third-party tool access compared to MCP-first agents
  • ❌ Privacy concerns around Meta's data practices
  • ❌ No code/API surface for developer extensibility

When to Use

  • Use when: You want a polished consumer assistant that lives across your phone and apps
  • Avoid when: You need programmatic extensibility (use Claude Code, Cursor, or MCP-first agents); you have strict privacy requirements around Meta's data practices

Resources